Semantic Attribution on EPIC-KITCHENS M (test)
84.91Average AccuracyMisFormer
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| MisFormerSubset evaluation (ChatGPT baseline fair comparison)=true2025.11 | 84.91 | 84.78 | 76.35 | 75.31 | 93.47 | 94.25 | |
| MisFormerSubset evaluation (ChatGPT baseline fair comparison)=false2025.11 | 84.13 | 83.89 | 76.83 | 76.43 | 91.43 | 91.34 | |
| ChatGPT2025.11 | 77.66 | 77.23 | 66.23 | 65.11 | 89.09 | 89.35 | |
| LLaVA-Vid2025.11 | 71.71 | 70.34 | 64.1 | 64.55 | 79.32 | 76.13 | |
| Vid-Chat2025.11 | 64.17 | 63.27 | 54.45 | 52.11 | 73.89 | 74.43 |